Quality of fingerprint scans captured using Optical Coherence Tomography

The performance limitations of the state-of-the-art methods for fingerprint presentation attack detection motivate the application of the Optical Coherence Tomography as the scanning technology capable of capturing all the information necessary for both the fingerprint identification and the presentation attack detection. The previous research has evidenced the need for a reliable technique for assessing quality of the captured scans so that the compliance of the capture subject behavior with acquisition rules could be verified. Good quality of these scans is not only a precondition of good system performance but it also enables reliable processing of the scanned data in the first place. This paper's contribution consists of a novel approach for estimation of the layered structure of an OCT fingerprint scan as a set of analytical surfaces and a subsequent analysis of the quality property of the obtained scan that is related to non-compliant behavior of the capture subjects. The method was able to detect 98% of low-quality scans at a false detection rate of 3%.